Theo Gevers conducts research in the field of computer vision, and in particular on the theories and methods that play a key role in enabling computers to analyse and understand images and videos. Among the techniques he uses colour analysis, given that colour is an important attribute for the recognition of objects. In analysing colours, colour constant and invariant representations are obtained for large scale visual search. Another focus area in Gevers's research is the study of human behaviour in videos. This research includes automatic recognition where a person is looking at their age, their posture and whether their laughter is genuine or fake. Facial expressions are used to automatically determine if people are happy, surprised or angry (emotion recognition). More recently, Gevers has been investigating the possibilities for 3D object reconstruction, involving scanning 3D objects (people, animals, tables, chairs) and spaces (room, apartments) and subsequently printing them using a 3D printer. Using newly developed algorithms, it is already possible to reconstruct 3D objects in real time. These 3D object models can then be used for the purposes of indexation, 3D object recognition and 3D printing.
Gevers is associate professor of Computer Science at the UvA and co-founder of Sightcorp bv, a UvA spin-off specialised in computer vision technology, facial analysis and emotion recognition technology. He is also the initiator of the company 3Duniversum, which supplies software for 3D object scanning and 3D printing. Gevers received a Vici grant from the Netherlands Organisation for Scientific Research (NWO) in 2007, and was previously ICREA Full Professor in Barcelona. He has published in numerous leading scientific journals, including IEEE Transactions on Pattern Analysis and Machine Intelligence, The International Journal in Computer Vision and IEEE Transactions on Image Processing, and is the author of the book Color in Computer Vision.